> From: Ian Clarke [mailto:I.Clarke at dynamicblue.com] > > If someone can provide a good case for why XML would be good for > metadata specifically (and "XML is a cool technology" is not a good > answer), then we will use it, but for the moment it strikes me as being > like using a nuclear weapon to swat a fly. >
Metadata can be generated and used by multiple layers of software, in and above Freenet. At a "rich client" level, like a browser, it makes much sense to use RDF, mostly because lots of other folks are doing exactly that and it would be foolish to ignore them and invent a different wheel. At the protocol level, you're way right, it doesn't make sense to use XML. Even a non-validating SAX1 parser is going to bloat the slim distribution; and XML will look nasty in what's currently a somewhat human-readable set of headers (especially considering it would have to be encoded to avoid CR-LF, yucky). The important question is: what metadata will be in the protocol? If the consensus answer to that question is "lots" (which I doubt) then MAYBE it makes sense to use XML. I vote for Content-Type and the rest of the MIME crew (MIME-Version, ...??) _______________________________________________ Freenet-dev mailing list Freenet-dev at lists.sourceforge.net http://lists.sourceforge.net/mailman/listinfo/freenet-dev
